[U-Boot] [PATCH v4 0/5] Implement "fastboot flash" for eMMC

This series implements the "fastboot flash" command for eMMC devices. It supports both raw and sparse images.
NOTES: - the support for the "fastboot flash" command is enabled with CONFIG_FASTBOOT_FLASH - the support for eMMC is enabled with CONFIG_FASTBOOT_FLASH_MMC_DEV - (future) the support for NAND would be enabled with CONFIG_FASTBOOT_FLASH_NAND(???)
This has been tested on ARMv7.
---- While we are struggling with the "sparse_format" copyright and licensing issues, can we accept the first three patches? Thanks, Steve
Changes in v4: - rearranged this patchset so that "sparse_format.h" can be dropped (if we cannot resolve the copyright/licensing issues) - update mmc_get_dev(...) to get_dev("mmc",....) - update printf() to puts() where applicable - update debug string as per feedback - rearranged "sparse format" support in this patchset, in order to isolate...
Changes in v3: - remove most references to 'mmc', which leaves only one mmc specific function: mmc_get_dev()
Changes in v2: - split large function into three - improved handling of response messages - additional partition size checking when writing sparse image - update README.android-fastboot file - new in v2
Steve Rae (5): usb/gadget: fastboot: add eMMC support for flash command usb/gadget: fastboot: add support for flash command usb/gadget: fastboot: minor cleanup usb/gadget: fastboot: add sparse image definitions usb/gadget: fastboot: implement sparse format
README | 10 +++ common/Makefile | 5 ++ common/cmd_fastboot.c | 7 +- common/fb_mmc.c | 191 ++++++++++++++++++++++++++++++++++++++++ doc/README.android-fastboot | 5 +- drivers/usb/gadget/f_fastboot.c | 44 ++++++++- include/fb_mmc.h | 8 ++ include/sparse_format.h | 58 ++++++++++++ 8 files changed, 319 insertions(+), 9 deletions(-) create mode 100644 common/fb_mmc.c create mode 100644 include/fb_mmc.h create mode 100644 include/sparse_format.h

- add support for 'fastboot flash' command for eMMC devices
Signed-off-by: Steve Rae srae@broadcom.com ---
Changes in v4: - rearranged this patchset so that "sparse_format.h" can be dropped (if we cannot resolve the copyright/licensing issues) - update mmc_get_dev(...) to get_dev("mmc",....) - update printf() to puts() where applicable
Changes in v3: - remove most references to 'mmc', which leaves only one mmc specific function: mmc_get_dev()
Changes in v2: - split large function into three - improved handling of response messages - additional partition size checking when writing sparse image
common/Makefile | 5 ++++ common/fb_mmc.c | 82 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++ include/fb_mmc.h | 8 ++++++ 3 files changed, 95 insertions(+) create mode 100644 common/fb_mmc.c create mode 100644 include/fb_mmc.h
diff --git a/common/Makefile b/common/Makefile index de5cce8..daebe39 100644 --- a/common/Makefile +++ b/common/Makefile @@ -266,4 +266,9 @@ obj-$(CONFIG_IO_TRACE) += iotrace.o obj-y += memsize.o obj-y += stdio.o
+# This option is not just y/n - it can have a numeric value +ifdef CONFIG_FASTBOOT_FLASH_MMC_DEV +obj-y += fb_mmc.o +endif + CFLAGS_env_embedded.o := -Wa,--no-warn -DENV_CRC=$(shell tools/envcrc 2>/dev/null) diff --git a/common/fb_mmc.c b/common/fb_mmc.c new file mode 100644 index 0000000..f42a115 --- /dev/null +++ b/common/fb_mmc.c @@ -0,0 +1,82 @@ +/* + * Copyright 2014 Broadcom Corporation. + * + * SPDX-License-Identifier: GPL-2.0+ + */ + +#include <common.h> +#include <fb_mmc.h> +#include <part.h> + +/* The 64 defined bytes plus \0 */ +#define RESPONSE_LEN (64 + 1) + +static char *response_str; + +static void fastboot_resp(const char *s) +{ + strncpy(response_str, s, RESPONSE_LEN); + response_str[RESPONSE_LEN - 1] = '\0'; +} + +static void write_raw_image(block_dev_desc_t *dev_desc, disk_partition_t *info, + const char *part_name, void *buffer, + unsigned int download_bytes) +{ + lbaint_t blkcnt; + lbaint_t blks; + + /* determine number of blocks to write */ + blkcnt = ((download_bytes + (info->blksz - 1)) & ~(info->blksz - 1)); + blkcnt = blkcnt / info->blksz; + + if (blkcnt > info->size) { + printf("%s: too large for partition: '%s'\n", __func__, + part_name); + fastboot_resp("FAILtoo large for partition"); + return; + } + + puts("Flashing Raw Image\n"); + + blks = dev_desc->block_write(dev_desc->dev, info->start, blkcnt, + buffer); + if (blks != blkcnt) { + printf("%s: failed writing to device %d\n", __func__, + dev_desc->dev); + fastboot_resp("FAILfailed writing to device"); + return; + } + + printf("........ wrote " LBAFU " bytes to '%s'\n", blkcnt * info->blksz, + part_name); + fastboot_resp("OKAY"); +} + +void fb_mmc_flash_write(const char *cmd, void *download_buffer, + unsigned int download_bytes, char *response) +{ + int ret; + block_dev_desc_t *dev_desc; + disk_partition_t info; + + /* initialize the response buffer */ + response_str = response; + + dev_desc = get_dev("mmc", CONFIG_FASTBOOT_FLASH_MMC_DEV); + if (!dev_desc || dev_desc->type == DEV_TYPE_UNKNOWN) { + printf("%s: invalid mmc device\n", __func__); + fastboot_resp("FAILinvalid mmc device"); + return; + } + + ret = get_partition_info_efi_by_name(dev_desc, cmd, &info); + if (ret) { + printf("%s: cannot find partition: '%s'\n", __func__, cmd); + fastboot_resp("FAILcannot find partition"); + return; + } + + write_raw_image(dev_desc, &info, cmd, download_buffer, + download_bytes); +} diff --git a/include/fb_mmc.h b/include/fb_mmc.h new file mode 100644 index 0000000..1ad1d13 --- /dev/null +++ b/include/fb_mmc.h @@ -0,0 +1,8 @@ +/* + * Copyright 2014 Broadcom Corporation. + * + * SPDX-License-Identifier: GPL-2.0+ + */ + +void fb_mmc_flash_write(const char *cmd, void *download_buffer, + unsigned int download_bytes, char *response);

- implement 'fastboot flash' for eMMC devices
Signed-off-by: Steve Rae srae@broadcom.com Reviewed-by: Marek Vasut marex@denx.de ---
Changes in v4: None Changes in v3: None Changes in v2: - update README.android-fastboot file
README | 10 ++++++++++ doc/README.android-fastboot | 5 +++-- drivers/usb/gadget/f_fastboot.c | 31 +++++++++++++++++++++++++++++++ 3 files changed, 44 insertions(+), 2 deletions(-)
diff --git a/README b/README index 1d71359..ed26884 100644 --- a/README +++ b/README @@ -1623,6 +1623,16 @@ The following options need to be configured: downloads. This buffer should be as large as possible for a platform. Define this to the size available RAM for fastboot.
+ CONFIG_FASTBOOT_FLASH + The fastboot protocol includes a "flash" command for writing + the downloaded image to a non-volatile storage device. Define + this to enable the "fastboot flash" command. + + CONFIG_FASTBOOT_FLASH_MMC_DEV + The fastboot "flash" command requires addition information + regarding the non-volatile storage device. Define this to + the eMMC device that fastboot should use to store the image. + - Journaling Flash filesystem support: CONFIG_JFFS2_NAND, CONFIG_JFFS2_NAND_OFF, CONFIG_JFFS2_NAND_SIZE, CONFIG_JFFS2_NAND_DEV diff --git a/doc/README.android-fastboot b/doc/README.android-fastboot index f1d128c..430e29c 100644 --- a/doc/README.android-fastboot +++ b/doc/README.android-fastboot @@ -6,8 +6,9 @@ Overview The protocol that is used over USB is described in README.android-fastboot-protocol in same directory.
-The current implementation does not yet support the flash and erase -commands. +The current implementation does not yet support the erase command or the +"oem format" command, and there is minimal support for the flash command; +it only supports eMMC devices.
Client installation =================== diff --git a/drivers/usb/gadget/f_fastboot.c b/drivers/usb/gadget/f_fastboot.c index 7a1acb9..e2659fa 100644 --- a/drivers/usb/gadget/f_fastboot.c +++ b/drivers/usb/gadget/f_fastboot.c @@ -19,6 +19,9 @@ #include <linux/compiler.h> #include <version.h> #include <g_dnl.h> +#ifdef CONFIG_FASTBOOT_FLASH_MMC_DEV +#include <fb_mmc.h> +#endif
#define FASTBOOT_VERSION "0.4"
@@ -469,6 +472,28 @@ static void cb_boot(struct usb_ep *ep, struct usb_request *req) fastboot_tx_write_str("OKAY"); }
+#ifdef CONFIG_FASTBOOT_FLASH +static void cb_flash(struct usb_ep *ep, struct usb_request *req) +{ + char *cmd = req->buf; + char response[RESPONSE_LEN]; + + strsep(&cmd, ":"); + if (!cmd) { + printf("%s: missing partition name\n", __func__); + fastboot_tx_write_str("FAILmissing partition name"); + return; + } + + strcpy(response, "FAILno flash device defined"); +#ifdef CONFIG_FASTBOOT_FLASH_MMC_DEV + fb_mmc_flash_write(cmd, (void *)CONFIG_USB_FASTBOOT_BUF_ADDR, + download_bytes, response); +#endif + fastboot_tx_write_str(response); +} +#endif + struct cmd_dispatch_info { char *cmd; void (*cb)(struct usb_ep *ep, struct usb_request *req); @@ -488,6 +513,12 @@ static const struct cmd_dispatch_info cmd_dispatch_info[] = { .cmd = "boot", .cb = cb_boot, }, +#ifdef CONFIG_FASTBOOT_FLASH + { + .cmd = "flash", + .cb = cb_flash, + }, +#endif };
static void rx_handler_command(struct usb_ep *ep, struct usb_request *req)

- update static function - additional debugging statements - update "fastboot command" information - add missing include file - update spelling
Signed-off-by: Steve Rae srae@broadcom.com ---
Changes in v4: - update debug string as per feedback
Changes in v3: None Changes in v2: - new in v2
common/cmd_fastboot.c | 7 ++++--- drivers/usb/gadget/f_fastboot.c | 13 +++++++++---- 2 files changed, 13 insertions(+), 7 deletions(-)
diff --git a/common/cmd_fastboot.c b/common/cmd_fastboot.c index 83fa7bd..909616d 100644 --- a/common/cmd_fastboot.c +++ b/common/cmd_fastboot.c @@ -30,7 +30,8 @@ static int do_fastboot(cmd_tbl_t *cmdtp, int flag, int argc, char *const argv[]) }
U_BOOT_CMD( - fastboot, 1, 1, do_fastboot, - "fastboot - enter USB Fastboot protocol", - "" + fastboot, 1, 0, do_fastboot, + "use USB Fastboot protocol", + "\n" + " - run as a fastboot usb device" ); diff --git a/drivers/usb/gadget/f_fastboot.c b/drivers/usb/gadget/f_fastboot.c index e2659fa..3b588a9 100644 --- a/drivers/usb/gadget/f_fastboot.c +++ b/drivers/usb/gadget/f_fastboot.c @@ -10,6 +10,7 @@ * * SPDX-License-Identifier: GPL-2.0+ */ +#include <config.h> #include <common.h> #include <errno.h> #include <malloc.h> @@ -41,7 +42,7 @@ struct f_fastboot { struct usb_function usb_function;
- /* IN/OUT EP's and correspoinding requests */ + /* IN/OUT EP's and corresponding requests */ struct usb_ep *in_ep, *out_ep; struct usb_request *in_req, *out_req; }; @@ -293,7 +294,7 @@ static int fastboot_add(struct usb_configuration *c) } DECLARE_GADGET_BIND_CALLBACK(usb_dnl_fastboot, fastboot_add);
-int fastboot_tx_write(const char *buffer, unsigned int buffer_size) +static int fastboot_tx_write(const char *buffer, unsigned int buffer_size) { struct usb_request *in_req = fastboot_func->in_req; int ret; @@ -341,6 +342,7 @@ static void cb_getvar(struct usb_ep *ep, struct usb_request *req)
strsep(&cmd, ":"); if (!cmd) { + printf("%s: missing variable\n", __func__); fastboot_tx_write_str("FAILmissing var"); return; } @@ -361,6 +363,7 @@ static void cb_getvar(struct usb_ep *ep, struct usb_request *req) else strcpy(response, "FAILValue not set"); } else { + printf("%s: unknown variable: %s\n", __func__, cmd); strcpy(response, "FAILVariable not implemented"); } fastboot_tx_write_str(response); @@ -534,10 +537,12 @@ static void rx_handler_command(struct usb_ep *ep, struct usb_request *req) } }
- if (!func_cb) + if (!func_cb) { + printf("%s: unknown command: %s\n", __func__, cmdbuf); fastboot_tx_write_str("FAILunknown command"); - else + } else { func_cb(ep, req); + }
if (req->status == 0) { *cmdbuf = '\0';

- to prepare for the support of fastboot sparse images
Signed-off-by: Steve Rae srae@broadcom.com --- This file is ASIS from: https://raw.githubusercontent.com/AOSB/android_system_core/master/libsparse/... (commit 28fa5bc347390480fe190294c6c385b6a9f0d68b) except for the __UBOOT__ conditional include.
Changes in v4: None Changes in v3: None Changes in v2: None
include/sparse_format.h | 58 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 58 insertions(+) create mode 100644 include/sparse_format.h
diff --git a/include/sparse_format.h b/include/sparse_format.h new file mode 100644 index 0000000..21fbd05 --- /dev/null +++ b/include/sparse_format.h @@ -0,0 +1,58 @@ +/* + * Copyright (C) 2010 The Android Open Source Project + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +#ifndef _LIBSPARSE_SPARSE_FORMAT_H_ +#define _LIBSPARSE_SPARSE_FORMAT_H_ +#define __UBOOT__ +#ifndef __UBOOT__ +#include "sparse_defs.h" +#endif + +typedef struct sparse_header { + __le32 magic; /* 0xed26ff3a */ + __le16 major_version; /* (0x1) - reject images with higher major versions */ + __le16 minor_version; /* (0x0) - allow images with higer minor versions */ + __le16 file_hdr_sz; /* 28 bytes for first revision of the file format */ + __le16 chunk_hdr_sz; /* 12 bytes for first revision of the file format */ + __le32 blk_sz; /* block size in bytes, must be a multiple of 4 (4096) */ + __le32 total_blks; /* total blocks in the non-sparse output image */ + __le32 total_chunks; /* total chunks in the sparse input image */ + __le32 image_checksum; /* CRC32 checksum of the original data, counting "don't care" */ + /* as 0. Standard 802.3 polynomial, use a Public Domain */ + /* table implementation */ +} sparse_header_t; + +#define SPARSE_HEADER_MAGIC 0xed26ff3a + +#define CHUNK_TYPE_RAW 0xCAC1 +#define CHUNK_TYPE_FILL 0xCAC2 +#define CHUNK_TYPE_DONT_CARE 0xCAC3 +#define CHUNK_TYPE_CRC32 0xCAC4 + +typedef struct chunk_header { + __le16 chunk_type; /* 0xCAC1 -> raw; 0xCAC2 -> fill; 0xCAC3 -> don't care */ + __le16 reserved1; + __le32 chunk_sz; /* in blocks in output image */ + __le32 total_sz; /* in bytes of chunk input file including chunk header and data */ +} chunk_header_t; + +/* Following a Raw or Fill or CRC32 chunk is data. + * For a Raw chunk, it's the data in chunk_sz * blk_sz. + * For a Fill chunk, it's 4 bytes of the fill data. + * For a CRC32 chunk, it's 4 bytes of CRC32 + */ + +#endif

On Thursday, August 07, 2014 at 01:55:12 AM, Steve Rae wrote:
- to prepare for the support of fastboot sparse images
Signed-off-by: Steve Rae srae@broadcom.com
Are we discussing the licensing issues here still ? Tom ?
Best regards, Marek Vasut

On 14-08-06 05:14 PM, Marek Vasut wrote:
On Thursday, August 07, 2014 at 01:55:12 AM, Steve Rae wrote:
- to prepare for the support of fastboot sparse images
Signed-off-by: Steve Rae srae@broadcom.com
Are we discussing the licensing issues here still ? Tom ?
Best regards, Marek Vasut
I hope so -- and I'm hoping that someone can help resolve this!!!! Thanks in advance, Steve

On Thursday, August 07, 2014 at 02:31:17 AM, Steve Rae wrote:
On 14-08-06 05:14 PM, Marek Vasut wrote:
On Thursday, August 07, 2014 at 01:55:12 AM, Steve Rae wrote:
- to prepare for the support of fastboot sparse images
Signed-off-by: Steve Rae srae@broadcom.com
Are we discussing the licensing issues here still ? Tom ?
Best regards, Marek Vasut
I hope so -- and I'm hoping that someone can help resolve this!!!! Thanks in advance, Steve
Someone should ask the legal dept ...
Best regards, Marek Vasut

Dear Marek,
In message 201408070214.35862.marex@denx.de you wrote:
Are we discussing the licensing issues here still ? Tom ?
Well, of course? We cannot add code that comes under an incompatible license. And we use SPDX ids.
Best regards,
Wolfgang Denk

On Thursday, August 07, 2014 at 07:36:49 AM, Wolfgang Denk wrote:
Dear Marek,
In message 201408070214.35862.marex@denx.de you wrote:
Are we discussing the licensing issues here still ? Tom ?
Well, of course? We cannot add code that comes under an incompatible license. And we use SPDX ids.
Well I agree Apache license is not compatible, but how can we proceed here for the good for U-Boot project and happiness of all involved parties here ? Aka. what do we do now to make the best of the contribution ?
Best regards, Marek Vasut

Dear Steve Rae,
In message 1407369313-13815-5-git-send-email-srae@broadcom.com you wrote:
- to prepare for the support of fastboot sparse images
Signed-off-by: Steve Rae srae@broadcom.com
This file is ASIS from: https://raw.githubusercontent.com/AOSB/android_system_core/master/libsparse/... (commit 28fa5bc347390480fe190294c6c385b6a9f0d68b) except for the __UBOOT__ conditional include.
Please make sure this origin documentation becomes part of the commit message - here in the comment section it gets lost. Please fix for all your patches.
- Copyright (C) 2010 The Android Open Source Project
- Licensed under the Apache License, Version 2.0 (the "License");
- you may not use this file except in compliance with the License.
- You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
- Unless required by applicable law or agreed to in writing, software
- distributed under the License is distributed on an "AS IS" BASIS,
- WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
- See the License for the specific language governing permissions and
- limitations under the License.
- */
Please always use SPDX license tags only.
Please note that this license is not compatible with GPL version 2, because it has some requirements that are not in that GPL version. An U-Boot is covered by GPLv2 ...
Best regards,
Wolfgang Denk

On Thursday, August 07, 2014 at 07:34:46 AM, Wolfgang Denk wrote:
Dear Steve Rae,
In message 1407369313-13815-5-git-send-email-srae@broadcom.com you wrote:
- to prepare for the support of fastboot sparse images
Signed-off-by: Steve Rae srae@broadcom.com
This file is ASIS from: https://raw.githubusercontent.com/AOSB/android_system_core/master/libsp arse/sparse_format.h (commit 28fa5bc347390480fe190294c6c385b6a9f0d68b)
except for the __UBOOT__ conditional include.
Please make sure this origin documentation becomes part of the commit message - here in the comment section it gets lost. Please fix for all your patches.
- Copyright (C) 2010 The Android Open Source Project
- Licensed under the Apache License, Version 2.0 (the "License");
- you may not use this file except in compliance with the License.
- You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
- Unless required by applicable law or agreed to in writing, software
- distributed under the License is distributed on an "AS IS" BASIS,
- WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
implied. + * See the License for the specific language governing permissions and + * limitations under the License.
- */
Please always use SPDX license tags only.
Please note that this license is not compatible with GPL version 2, because it has some requirements that are not in that GPL version. An U-Boot is covered by GPLv2 ...
I think we should start considering getting some kind of a legal conseling here, or do we actually have one already ?
Best regards, Marek Vasut

On 08/07/2014 03:09 PM, Marek Vasut wrote:
- Copyright (C) 2010 The Android Open Source Project
- Licensed under the Apache License, Version 2.0 (the "License");
Please note that this license is not compatible with GPL version 2, because it has some requirements that are not in that GPL version. An U-Boot is covered by GPLv2 ...
I think we should start considering getting some kind of a legal conseling here, or do we actually have one already ?
I used the gnu page back then. Apache2 is compatible with GPLv3 but not with GPLv2 [0]. Some parts of the code I played around (the android image header file for instance) was under the Apache2 license and in another git tree under the two-clause BSD license which was compatible with GPLv2 [1]. So the argument about incompatible licenses is valid here. Should u-boot switch to GPLv3 you need to hunt down that file again and this time with the apache2 license :)
[0] https://www.gnu.org/licenses/license-list.html#apache2 [1] https://www.gnu.org/licenses/license-list.html#FreeBSD
Best regards, Marek Vasut
Sebastian

On Thursday, August 07, 2014 at 03:29:17 PM, Sebastian Andrzej Siewior wrote:
On 08/07/2014 03:09 PM, Marek Vasut wrote:
- Copyright (C) 2010 The Android Open Source Project
- Licensed under the Apache License, Version 2.0 (the "License");
Please note that this license is not compatible with GPL version 2, because it has some requirements that are not in that GPL version. An U-Boot is covered by GPLv2 ...
I think we should start considering getting some kind of a legal conseling here, or do we actually have one already ?
I used the gnu page back then. Apache2 is compatible with GPLv3 but not with GPLv2 [0]. Some parts of the code I played around (the android image header file for instance) was under the Apache2 license and in another git tree under the two-clause BSD license which was compatible with GPLv2 [1]. So the argument about incompatible licenses is valid here. Should u-boot switch to GPLv3 you need to hunt down that file again and this time with the apache2 license :)
I agree. Is there a version of this file with BSD license on it ? Then that'd be a problem solved, right ? The other option would be to ask Google to release it under BSD license, but I don't know how easy that would be.
Best regards, Marek Vasut

On 08/07/2014 03:45 PM, Marek Vasut wrote:
I agree. Is there a version of this file with BSD license on it ? Then that'd be a problem solved, right ?
Right. I don't recall the locations of the second tree but if Steve is going to search it, he should find it.
Best regards, Marek Vasut
Sebastian

On Thursday, August 07, 2014 at 03:51:54 PM, Sebastian Andrzej Siewior wrote:
On 08/07/2014 03:45 PM, Marek Vasut wrote:
I agree. Is there a version of this file with BSD license on it ? Then that'd be a problem solved, right ?
Right. I don't recall the locations of the second tree but if Steve is going to search it, he should find it.
Then we're good here, thanks for pointing this out !
Best regards, Marek Vasut

On 14-08-07 06:51 AM, Sebastian Andrzej Siewior wrote:
On 08/07/2014 03:45 PM, Marek Vasut wrote:
I agree. Is there a version of this file with BSD license on it ? Then that'd be a problem solved, right ?
Right. I don't recall the locations of the second tree but if Steve is going to search it, he should find it.
I have not found anything other than the Apache 2 license (yet). If anyone can find anything, please let me know! Thanks in advance, Steve
Best regards, Marek Vasut
Sebastian

Dear Sebastian,
In message 53E37F2D.5090700@linutronix.de you wrote:
I think we should start considering getting some kind of a legal conseling here, or do we actually have one already ?
I used the gnu page back then. Apache2 is compatible with GPLv3 but not with GPLv2 [0]. Some parts of the code I played around (the android image header file for instance) was under the Apache2 license and in another git tree under the two-clause BSD license which was compatible with GPLv2 [1].
We definitely need GPLv2 compatibility.
So the argument about incompatible licenses is valid here. Should u-boot switch to GPLv3 you need to hunt down that file again and this
You would have to rewrite the major part of U-Boot or to cripple it to uselessness if you try to release it under GPLv3. I cannot see anybody investing any efforts in that direction; all these threads about a potential GPLv3 switch are just a hollow scarecrow.
It seems clear to me that GPLv3 is not an acceptable solution for developers, vendors, and end users simultanously. When I ask for GPLv2+ I hope that some later version of GPL may find better general support.
Best regards,
Wolfgang Denk

On 07-08-14 22:59, Wolfgang Denk wrote:
<snip> > So the argument about incompatible licenses is valid here. Should > u-boot switch to GPLv3 you need to hunt down that file again and this You would have to rewrite the major part of U-Boot or to cripple it
noise ^ not
to uselessness if you try to release it under GPLv3. I cannot see anybody investing any efforts in that direction; all these threads about a potential GPLv3 switch are just a hollow scarecrow.
It seems clear to me that GPLv3 is not an acceptable solution for developers, vendors, and end users simultanously. When I ask for GPLv2+ I hope that some later version of GPL may find better general support.
Regards, Jeroen

Dear Jeroen,
In message 53E3E974.1020504@myspectrum.nl you wrote:
On 07-08-14 22:59, Wolfgang Denk wrote:
<snip> > So the argument about incompatible licenses is valid here. Should > u-boot switch to GPLv3 you need to hunt down that file again and this You would have to rewrite the major part of U-Boot or to cripple it
noise ^ not
to uselessness if you try to release it under GPLv3. I cannot see anybody investing any efforts in that direction; all these threads about a potential GPLv3 switch are just a hollow scarecrow.
sorry, can't parse that.
Best regards,
Wolfgang Denk

On 07-08-14 23:10, Wolfgang Denk wrote:
Dear Jeroen,
In message 53E3E974.1020504@myspectrum.nl you wrote:
On 07-08-14 22:59, Wolfgang Denk wrote:
<snip> > So the argument about incompatible licenses is valid here. Should > u-boot switch to GPLv3 you need to hunt down that file again and this You would have to rewrite the major part of U-Boot or to cripple it
noise ^ not
to uselessness if you try to release it under GPLv3. I cannot see anybody investing any efforts in that direction; all these threads about a potential GPLv3 switch are just a hollow scarecrow.
sorry, can't parse that.
no, this was nicely formatted in my mail editor. Just wanted to point out you forgot a "not". You mentioned "to cripple it" you more then like ment "to _not_ cripple it". But the whitespace got lost along the way.
Regards, Jeroen

On 07-08-14 23:15, Jeroen Hofstee wrote:
On 07-08-14 23:10, Wolfgang Denk wrote:
Dear Jeroen,
In message 53E3E974.1020504@myspectrum.nl you wrote:
On 07-08-14 22:59, Wolfgang Denk wrote:
<snip> > So the argument about incompatible licenses is valid here. Should > u-boot switch to GPLv3 you need to hunt down that file again and this You would have to rewrite the major part of U-Boot or to cripple it
noise ^ not
to uselessness if you try to release it under GPLv3. I cannot see anybody investing any efforts in that direction; all these threads about a potential GPLv3 switch are just a hollow scarecrow.
sorry, can't parse that.
no, this was nicely formatted in my mail editor. Just wanted to point out you forgot a "not". You mentioned "to cripple it" you more then like ment "to _not_ cripple it". But the whitespace got lost along the way.
ah, and the noise: part is that most readers would have figured that out without me mentioning it.
Regards, Jeroen

Dear Jeroen,
In message 53E3EC54.4090002@myspectrum.nl you wrote:
So the argument about incompatible licenses is valid here. Should u-boot switch to GPLv3 you need to hunt down that file again and this
You would have to rewrite the major part of U-Boot or to cripple it
noise ^ not
to uselessness if you try to release it under GPLv3. I cannot see anybody investing any efforts in that direction; all these threads about a potential GPLv3 switch are just a hollow scarecrow.
sorry, can't parse that.
no, this was nicely formatted in my mail editor. Just wanted to point out you forgot a "not". You mentioned "to cripple it" you more then like ment "to _not_ cripple it". But the whitespace got lost along the way.
No, I did NOT forget a "not".
There is an awful lot of GPLv2 (V2 only) code in U-Boot. Many of the most interesting features thaty have been imported from the Linux kernel are GPLv2 only. Either you have to rewrite all this code. I don't see any volunteers for that in the community, and neither I see any commercial interest for that. So alternatively you would have to remove resp. disable in the configuration all such code - which would cripple U-Boot, as it would remove a lot of really useful functionali- ty, rendering it more or less useless for many practical use cases.
All people claiming "U-Boot will go GPLv3 tomorrow" are just spreading FUD.
Best regards,
Wolfgang Denk

On Thursday, August 07, 2014 at 11:36:50 PM, Wolfgang Denk wrote:
Dear Jeroen,
In message 53E3EC54.4090002@myspectrum.nl you wrote:
So the argument about incompatible licenses is valid here. Should u-boot switch to GPLv3 you need to hunt down that file again and this
You would have to rewrite the major part of U-Boot or to cripple it
noise ^ not
to uselessness if you try to release it under GPLv3. I cannot see anybody investing any efforts in that direction; all these threads about a potential GPLv3 switch are just a hollow scarecrow.
sorry, can't parse that.
no, this was nicely formatted in my mail editor. Just wanted to point out you forgot a "not". You mentioned "to cripple it" you more then like ment "to _not_ cripple it". But the whitespace got lost along the way.
No, I did NOT forget a "not".
There is an awful lot of GPLv2 (V2 only) code in U-Boot.
OK, so we need a GPLv2-compatible header file. That's the conclusion.
Can we get one or how shall we proceed in the opposite case please?
Best regards, Marek Vasut

Dear Marek,
In message 201408072343.58591.marex@denx.de you wrote:
OK, so we need a GPLv2-compatible header file. That's the conclusion.
Well, GPLv2 compatible might be acceptable, but GPLv2+ compatible woould be best.
Best regards,
Wolfgang Denk

On Friday, August 08, 2014 at 12:17:38 AM, Wolfgang Denk wrote:
Dear Marek,
In message 201408072343.58591.marex@denx.de you wrote:
OK, so we need a GPLv2-compatible header file. That's the conclusion.
Well, GPLv2 compatible might be acceptable, but GPLv2+ compatible woould be best.
I was talking to Hans about this yesterday (and who was dropped from the CC for some reason) and it turns out the easy way out here is to just reimplement the header. I would like him to take a look and express his opinion properly though.
Best regards, Marek Vasut

( Google agreed to submit a patch with BSD3 license.... stay tuned!!! )
On 14-08-09 05:48 AM, Marek Vasut wrote:
On Friday, August 08, 2014 at 12:17:38 AM, Wolfgang Denk wrote:
Dear Marek,
In message 201408072343.58591.marex@denx.de you wrote:
OK, so we need a GPLv2-compatible header file. That's the conclusion.
Well, GPLv2 compatible might be acceptable, but GPLv2+ compatible woould be best.
I was talking to Hans about this yesterday (and who was dropped from the CC for some reason) and it turns out the easy way out here is to just reimplement the header. I would like him to take a look and express his opinion properly though.
Best regards, Marek Vasut

On Saturday, August 09, 2014 at 06:53:24 PM, Steve Rae wrote:
( Google agreed to submit a patch with BSD3 license.... stay tuned!!! )
Thank you for your work and perseverance, really !
Best regards, Marek Vasut

On Thursday, August 07, 2014 at 11:15:00 PM, Jeroen Hofstee wrote:
On 07-08-14 23:10, Wolfgang Denk wrote:
Dear Jeroen,
In message 53E3E974.1020504@myspectrum.nl you wrote:
On 07-08-14 22:59, Wolfgang Denk wrote:
<snip>
So the argument about incompatible licenses is valid here. Should u-boot switch to GPLv3 you need to hunt down that file again and this
You would have to rewrite the major part of U-Boot or to cripple it
noise ^ not
to uselessness if you try to release it under GPLv3. I cannot see anybody investing any efforts in that direction; all these threads about a potential GPLv3 switch are just a hollow scarecrow.
sorry, can't parse that.
no, this was nicely formatted in my mail editor. Just wanted to point out you forgot a "not". You mentioned "to cripple it" you more then like ment "to _not_ cripple it". But the whitespace got lost along the way.
Please be more explicit next time, really. It can be difficult to find the balance between keeping the mail short and being explicit, but usually it's sufficient to read the mail after a few minutes and see if you'd understand the message yourself if you got one as such.
Thanks!
Best regards, Marek Vasut

- add capability to "fastboot flash" with sparse format images
Signed-off-by: Steve Rae srae@broadcom.com --- I suspect that the "sparse image" handling (ie. the "while (remaining_chunks)" loop) has been implemented elsewhere -- I need help finding the original code to determine any licensing issues.... Thanks, Steve
Changes in v4: - rearranged "sparse format" support in this patchset, in order to isolate...
Changes in v3: None Changes in v2: None
common/fb_mmc.c | 115 ++++++++++++++++++++++++++++++++++++++++++++++++++++++-- 1 file changed, 112 insertions(+), 3 deletions(-)
diff --git a/common/fb_mmc.c b/common/fb_mmc.c index f42a115..306c102 100644 --- a/common/fb_mmc.c +++ b/common/fb_mmc.c @@ -1,5 +1,6 @@ /* - * Copyright 2014 Broadcom Corporation. + * Copyright TODO + * Portions Copyright 2014 Broadcom Corporation. * * SPDX-License-Identifier: GPL-2.0+ */ @@ -7,6 +8,7 @@ #include <common.h> #include <fb_mmc.h> #include <part.h> +#include <sparse_format.h>
/* The 64 defined bytes plus \0 */ #define RESPONSE_LEN (64 + 1) @@ -19,6 +21,108 @@ static void fastboot_resp(const char *s) response_str[RESPONSE_LEN - 1] = '\0'; }
+static int is_sparse_image(void *buf) +{ + sparse_header_t *s_header = (sparse_header_t *)buf; + + if ((le32_to_cpu(s_header->magic) == SPARSE_HEADER_MAGIC) && + (le16_to_cpu(s_header->major_version) == 1)) + return 1; + + return 0; +} + +static void write_sparse_image(block_dev_desc_t *dev_desc, + disk_partition_t *info, const char *part_name, + void *buffer, unsigned int download_bytes) +{ + lbaint_t blk; + lbaint_t blkcnt; + lbaint_t blks; + sparse_header_t *s_header = (sparse_header_t *)buffer; + chunk_header_t *c_header; + void *buf; + uint32_t blk_sz; + uint32_t remaining_chunks; + uint32_t bytes_written = 0; + + blk_sz = le32_to_cpu(s_header->blk_sz); + + /* verify s_header->blk_sz is exact multiple of info->blksz */ + if (blk_sz != (blk_sz & ~(info->blksz - 1))) { + printf("%s: Sparse image block size issue [%u]\n", + __func__, blk_sz); + fastboot_resp("FAILsparse image block size issue"); + return; + } + + if ((le32_to_cpu(s_header->total_blks) * blk_sz) > + (info->size * info->blksz)) { + printf("%s: Sparse image is too large for the partition\n", + __func__); + fastboot_resp("FAILsparse image is too large"); + return; + } + + puts("Flashing Sparse Image\n"); + + remaining_chunks = le32_to_cpu(s_header->total_chunks); + c_header = (chunk_header_t *)(buffer + + le16_to_cpu(s_header->file_hdr_sz)); + blk = info->start; + while (remaining_chunks) { + blkcnt = + (le32_to_cpu(c_header->chunk_sz) * blk_sz) / info->blksz; + + switch (le16_to_cpu(c_header->chunk_type)) { + case CHUNK_TYPE_RAW: + buf = (void *)c_header + + le16_to_cpu(s_header->chunk_hdr_sz); + + if (blk + blkcnt > info->start + info->size) { + printf( + "%s: Request would exceed partition size!\n", + __func__); + fastboot_resp( + "FAILRequest would exceed partition size!"); + return; + } + + blks = dev_desc->block_write(dev_desc->dev, blk, blkcnt, + buf); + if (blks != blkcnt) { + printf("%s: Write failed " LBAFU "\n", + __func__, blks); + fastboot_resp("FAILwrite failure"); + return; + } + + bytes_written += blkcnt * info->blksz; + break; + + case CHUNK_TYPE_FILL: + case CHUNK_TYPE_DONT_CARE: + case CHUNK_TYPE_CRC32: + /* do nothing */ + break; + + default: + /* error */ + printf("%s: Unknown chunk type\n", __func__); + fastboot_resp("FAILunknown chunk type in sparse image"); + return; + } + + blk += blkcnt; + c_header = (chunk_header_t *)((void *)c_header + + le32_to_cpu(c_header->total_sz)); + remaining_chunks--; + } + + printf("........ wrote %u bytes to '%s'\n", bytes_written, part_name); + fastboot_resp("OKAY"); +} + static void write_raw_image(block_dev_desc_t *dev_desc, disk_partition_t *info, const char *part_name, void *buffer, unsigned int download_bytes) @@ -77,6 +181,11 @@ void fb_mmc_flash_write(const char *cmd, void *download_buffer, return; }
- write_raw_image(dev_desc, &info, cmd, download_buffer, - download_bytes); + if (is_sparse_image(download_buffer)) { + write_sparse_image(dev_desc, &info, cmd, download_buffer, + download_bytes); + } else { + write_raw_image(dev_desc, &info, cmd, download_buffer, + download_bytes); + } }

On Thursday, August 07, 2014 at 01:55:08 AM, Steve Rae wrote:
This series implements the "fastboot flash" command for eMMC devices. It supports both raw and sparse images.
NOTES:
- the support for the "fastboot flash" command is enabled with
CONFIG_FASTBOOT_FLASH - the support for eMMC is enabled with CONFIG_FASTBOOT_FLASH_MMC_DEV - (future) the support for NAND would be enabled with CONFIG_FASTBOOT_FLASH_NAND(???)
This has been tested on ARMv7.
I'll just wait for Tom's comments on 4/5 and 5/5 and will see then. The 1-3/5 look OK.
Best regards, Marek Vasut

On 14-08-06 05:16 PM, Marek Vasut wrote:
On Thursday, August 07, 2014 at 01:55:08 AM, Steve Rae wrote:
This series implements the "fastboot flash" command for eMMC devices. It supports both raw and sparse images.
NOTES:
- the support for the "fastboot flash" command is enabled with
CONFIG_FASTBOOT_FLASH - the support for eMMC is enabled with CONFIG_FASTBOOT_FLASH_MMC_DEV - (future) the support for NAND would be enabled with CONFIG_FASTBOOT_FLASH_NAND(???)
This has been tested on ARMv7.
I'll just wait for Tom's comments on 4/5 and 5/5 and will see then. The 1-3/5 look OK.
Best regards, Marek Vasut
THANKS! Steve
participants (5)
-
Jeroen Hofstee
-
Marek Vasut
-
Sebastian Andrzej Siewior
-
Steve Rae
-
Wolfgang Denk